Reasoning failures highlighted by Apple research on LLMs - eviltoast
  • rottingleaf@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    2
    ·
    17 days ago

    Because it’s not AI, it’s sophisticated pattern separation, recognition, lossy compression and extrapolation systems.

    Artificial intelligence, like any intelligence, has goals and priorities. It has positive and negative reinforcements from real inputs.

    Their AI will be possible when it’ll be able to want something and decide something, with that moment based on entropy and not extrapolation.

    • InternetPerson@lemmings.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      16 days ago

      Artificial intelligence, like any intelligence, has goals and priorities

      No. Intelligence does not necessitate goals. You are able to understand math, letters, words, meaning of those without pursuing a specific goal.

      Because it’s not AI, it’s sophisticated pattern separation, recognition, lossy compression and extrapolation systems.

      And our brains work in a similar way.

      • rottingleaf@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        1
        ·
        16 days ago

        Our brains work in various ways. Somewhere in there a system similar to those “AI”'s exists, I agree. It’s just only one part. Artificial dicks are not the same thing as artificial humans